PROFILING OF URINARY PROTEINS OF PATIENTS WITH PCa AND THOSE WITH BPH AS WELL AS CONTROLS

The second part of the current study involves experimental-designs mirroring the initial proteomic profiling of serum samples but on optimised concentration of acetone precipitated urinary proteins from healthy subjects, patients with BPH and those with PCa. Urine, albeit containing 1000-fold lower in protein concentration compared to plasma or serum, represents most of the plasma proteins. It also has increased proportions of low molecular weight protein and peptide components particularly those of the urinary tract (Hortin and Sviridov, 2007). In addition, the urinary proteome is also known to alter as results of pathophysiological conditions, and as such, provides an ideal alternative to plasma or serum for analysis (Veenstra et al., 2005). In addition to this, urine sample was also selected for its attractive features of non-invasiveness, ease of attainability and cost-effective sampling method (Pisitkun et al., 2006). The adaptation of acetone precipitation method on the other hand, had resulted in concentrated and purified urinary protein contents (Lovrien and Matulis, 1997) which was intended for the 2-DE proteomic profiling. The inclusion of age-matched healthy subjects as ‘non- cancer’ controls and a slight increase in the sample size in this second part of the study had favourably enhance the specificity as well as sensitivity of the potentially identified biomarkers.

This part of the study was initiated by subjecting an optimised concentration of acetone precipitated urinary proteins (100 µg) from patients with BPH and those with PCa as well as control subjects to 2-DE and silver staining, to investigate the differential levels of altered abundance of urinary proteins. The 2-DE resolved silver-stained urinary protein profiles of patients with BPH and those with PCa and control subjects were generally comparable with those detected in urine of patients with endometrial

University

cancer (Mu et al., 2012) but with little loss in proteins such as immunoglobulin gamma chains and KNG-1. The adaptation of acetone precipitation to the pre-processing of the urine samples had very likely contributed to the loss due to less than complete recovery of proteins (Puchades et al., 1999; Srivastava and Srivastava, 1998; Thongboonkerd et

al., 2002). Although, the levels of acetone (four parts of cold acetone) employed in this

study were generally acceptable (Botelho et al., 2010; Davidsson et al., 1999), but may not have been optimal for some proteins (Srivastava and Srivastava, 1998; Thongboonkerd et al., 2002). Nevertheless, incorporation of this method had certainly improved resolution and consistency of the 2-DE gels (Jayapalan et al., 2013; this thesis).

When comparative densitometric analysis was performed on the 2-DE resolved urinary protein profiles of patients and control subjects, different altered levels of abundance of three proteins were observed. Significantly lower abundance of PSAPf, AMBPf1 and AMBPf2 was demonstrated in the urine of PCa patients compared to control subjects. Abundance of AMBPf2, on the other hand, was also significantly different when comparison was made between 2-DE profiles generated from urine of patients with PCa and those with BPH.

The identities of proteins of altered abundance, including PSAPf, AMBPf1 and AMBPf2 were confirmed by MS. A detailed study of MS/MS-derived amino acid sequences of both urinary AMBPf1 and AMBPf2 that were detected in the present study indicates extensive sequence homology to inter-alpha-trypsin inhibitor light chain (ITIL) region of AMBP (amino acids 206 - 352). Four of the peptide hit sequences appeared in both AMBPf1 and AMBPf2 cluster of spots that were analysed. This, together with their differences of molecular weights and pIs, suggests that urinary AMBPf1 and AMBPf2 were the ITIL products of the different proteolytic processing of

University

AMBP. As such, urinary AMBPf1 and AMBPf2 were henceforth, referred as ITILf1 and ITILf2, respectively for convenience of ensuing discussion.

PSAP is a 68 - 73 kDa lysosomal compartmental precursor glycoprotein (Kishimoto et al., 1992), ubiquitously present either as secretory or integral membrane proteins in secretory fluids (e.g.: cerebrospinal fluid, seminal plasma, human milk, pancreatic juice, blood plasma and bile) (Hineno et al., 1991) and tissues (e.g., brain, testes, kidney, spleen and liver) (O'Brien et al., 1988). Initially expressed as a single precursor PSAP, all four saposins A, B, C and D are by-products of proteolytic processing of PSAP within the lysosomes with diverse biological involvement. The urinary PSAPf spot detected in the present study was most likely the saposin B peptides cleaved from the prosaposin precursor (Kishimoto et al., 1992). This was deduced based on the observed experimental mass and MS/MS-derived amino acid sequences of the protein (Table 4.4).

Aside from exerting essential neurotrophic and myelinotrophic activities (Misasi et al., 2009), PSAP is also a well-known pleiotropic growth factor (Wu et al., 2012). Overexpression of PSAP in breast cancer cell lines was thought to promote tumour growth in breast cancer by stimulating the oestrogen receptor alpha-mediated signalling axis (Wu et al., 2012). Similarly, amplified levels of PSAP were also demonstrated to promote carcinogenesis and progression of the prostate but through the activation of multiple signal transduction pathways and anti-apoptotic effect in metastatic androgen-independent prostate cancer cell lines (Koochekpour et al., 2005). In another study, influence of up-regulation of androgen receptors, PSA expression and cellular activity in androgen-sensitive human prostate adenocarcinoma cells were also thought to contribute to PCa (Koochekpour et al., 2007). Conversely, down-regulation of PSAP was shown to significantly decrease adhesion, migration and invasion of

University

metastatic prostate cancer cells via lysosomal proteolysis-dependent pathway by RNA interference (Hu et al., 2010). Although the mechanism is poorly understood, it was believed that the decreased level of urinary saposin B of PCa patients observed in this study may be related to the reduced levels of proteolytic processing of PSAP in the patients compared to control subjects.

ITIL, also known as bikunin, a split-product of proteolytically cleaved AMBP precursor containing two domains of Kunitz-type proteinase inhibitor with a theoretical Mr of 25 - 26 kDa, is released in free-form into the circulation (Fries and Blom, 2000;

Gebhard et al., 1990). Generally, the free-form bikunin is filtered in the glomeruli of the kidney and reabsorbed in the tubuli. However, due to physiological or pathological duress, the levels of reabsorption were thought to be compromised hence, the apparent high levels of bikunin in the urine (Blom et al., 1997). Additionally, these high levels were also hypothesised to have been contributed by the secretion of free-form bikunin from the tumour cells itself, in cancer (Hochstrasser et al., 1989).

Nevertheless, this is not the first time the long-term association of bikunin as being the major component of cancer-associated proteinuria (Chawla et al., 1992; Chawla et al., 1984) has been thwarted. To the best of knowledge, the present study was the first to report on the lower levels of ITIL peptides in the urine of patients with PCa compared to controls, particularly in terms of its diagnostic potential (Jayapalan et

al., 2013; this thesis). Aside from this, the reduced abundance of AMBPf2 peptide

observed in urine of patients with PCa compared to those with BPH, does offer a new dimension to the specificity of the identified potential marker.

As observed in this study, down-regulation of bikunin in urine of patients with bladder cancer (Tsui et al., 2010) has been previously reported as well. The mechanism of action of bikunin in cancer prevention lies in the suppression of various

University

signalling cascades (Kobayashi et al., 2001; Matsuzaki et al., 2004). As such, increased level of bikunin has been previously shown to prevent progression of ovarian cancer (Suzuki et al., 2003; Tanaka et al., 2003) by inhibiting expression of tumour cell- associated plasmin activity and urolinase-plasminogen activator expression at the levels of gene and protein (Kobayashi et al., 2000; Kobayashi et al., 2001) in ovarian carcinoma cell line. Therefore, the lower levels of bikunin detected in urine of patients with PCa in the present study possibly reflects localised cancer within the prostate gland thus, in line with the stages of cancer of the recruited PCa patients (stage I and II).

In the second approach, the 2-DE electrophoresed gels were electrotransferred onto a NC membrane for the detection O-glycosylated urinary proteins using HRP- conjugated CGB lectin. By subjecting the urinary O-glycosylated protein resolved on the lectin-detected NC membranes to densitometry analysis, significantly enhanced level of ITIH4f (at approximately 50 kDa) was noted in the urine of PCa patients compared to control subjects. No tangible differences in levels of other O-glycosylated urinary proteins were detected between patients with BPH and healthy controls. Despite many attempts, the identity of urinary ITIH4f on the NC membranes was unable to be confirmed using on-membrane trypsin digestion technique (Luque-Garcia and Neubert, 2009; Mu et al., 2012). Because of that, ITIH4f was initially determined by visual comparison to that of the optimised acetone precipitated urinary protein profiles. Later,

O-glycosylated urinary proteins was first isolated using CGB lectin affinity

chromatography, trypsin digested (in-solution) and finally, the presence of ITIH4f in urine was confirmed by nano LC-MS/MS.

Contrary to the ITIL peptides, the lectin-detected ITIH4f appeared to be overexpressed in the urine of patients with PCa as opposed to the controls and patients with BPH. ITIH4, a single chain acute-phase glycoprotein with an apparent Mr of 120-

University

kDa belongs to the inter-α-trypsin inhibitor (ITI) family of serine protease inhibitors (Fujita et al., 2004). Unlike its other ITIH family members, ITIH4 does not interact with ITIL due to the lack of a C-terminus region peptide sequence that is required for ITIL binding (Nishimura et al., 1995; Salier et al., 1996). Interestingly, ITIH, which is typically linked to ITIL via estrified chondroitin 4-sulfate, has been shown to be down- regulated in multiple solid tumours (Hamm et al., 2008), and this is also in agreement with the lower level of urinary ITIL that was detected in patients with PCa in the present study. Another unique property of ITIH4 is the kallikrein-released bradykinin-like domain perched on its C-terminal end (Nishimura et al., 1995), which in turns, allows fragmentation of the 120 kDa ITIH4 native protein by plasma kallikrein (Saguchi et al., 1995). The resulting fragmented patterns of 85 kDa and 35 kDa ITIH4fs have been invariably associated with various diseased conditions including cancer (Song et al., 2006), inflammation (Pineiro et al., 2004), normal and hydatiform molar pregnancy (Mohamed et al., 2013), amyotrophic lateral sclerosis (Tanaka et al., 2013), acute ischaemic stroke (Kashyap et al., 2009) and many others. When referred to the earlier findings of the present study, data on elevated levels of ITIH4f in the urine of patients with PCa neatly corroborates to the low levels of the serum ITIH4f in PCa patients as substantial amount of the peptide may have been excreted in the urine in order to clear the excess breakdown product of the ITIH4, which is the 35 kDa ITIH4f in the sera. Interestingly, the data of the present study also appear to be in direct contrast to what was previously reported for ovarian carcinoma, in which the ITIH4f was significantly enhanced in the patients’ sera (Mohamed et al., 2008) but lowered in abundance in the their urine (Abdullah-Soheimi et al., 2010).

University

5.4. POTENTIAL USE OF INDICES RELATING TO LEVELS OF SERUM